Probing the Magnetism and chemistry of a burried oxide layer.

Recent years have witnessed a tremendous growth of research in the field of magnetoelectronics, in view of its obvious potential for novel devices with entirely new capabilities. In this context, phenomena such as giant magnetoresistance (GMR), colossal magnetoresistance (CMR), spin-tunneling in junctions (STJ), and more recently spin coherence and spin phasing have attracted significant attention. The heterostructures/multilayers used in these studies generally involve ternary or quaternary alloys/compounds having magnetic, semiconducting, or superconducting properties, and the device action of interest generally occurs at the burried interface. It is thus of critical importance to know the properties of such an interface. Recetly, Idzerda and coworkers at NRL have developed a method to probe the burried interface, based on a combined use of X-ray absorption spectroscopy (XAS) and X-ray magnetic circular dichroism(X-MCD). In our work, done in collaboration with Idzerda, we have used this method to elementally map the magnetic and chemical quality of the interface of La_0.7Sr_0.3MnO₃ (LSMO) covered by various overlayers. LSMO is characteristic of the class of Mn perovskites which display an extremely high degree of spin polarization, making them preferred candidates for magnetic-based devices which exploit spin polarized electron transport.

XMCD is an element-specific, magnetic spectroscopic tool where the difference in the absorption of left- and right-circular polarized photons is measured at the absorption edges of the relevant elements. Two total electron yield spectra, one with the incident light helicity (circular polarization) oriented parallel to the remnant magnetization of the sample and one anti-parallel, are collected as the X-ray energy is swept continuously through the L2,3 edge energies of a constituent element. The XAS is the average of these two recorded spectra, while the XMCD is the difference between the two. In XAS and XMCD, the excited electron probes the unfilled states above the fermi level to reveal local chemical (and, also from XMCD, magnetic) information about the element. All data reported here were collected at the NRL-NSLS Magnetic Circular Dichroism Facility located at beamline U4B of the National Synchrotron Light Source (NSLS). With our experimental configuration, the measurements are sensitive to the top 50-100 Å of the LSMO film. Based on the XAS and XMCD data thus obtained we have been able to establish that the deposition of YBa₂Cu₃O_7-ò on La_0.7Sr_0.3MnO₃ results in the diffusion of La away from the interfacial region, strongly modifying the interfacial electronic and magnetic properties. The results for other overlayers such as SrTiO₃ or LaAlO₃ are different.

In the XAS spectra shown in fig. 1 for YBa₂Cu₃O_7-ò on La_0.7Sr_0.3MnO₃, there is a continuous energy shift of the L₃ peak, a narrowing in the energy distribution of the L₃ peak, and the development of a new feature at a fixed lower energy. Simultaneously, the XMCD intensity is observed to simply decrease with increasing YBCO coverage without any significant changes in the spectra. The variations in the XAS spectra with increased YBCO thickness are very similar to those for La_1-xSr_xMnO₃ as the composition is changed from La_0.7Sr_0.3MnO₃ to La_0.1Sr_0.9MnO₃. This strongly suggests that the deposition of YBCO on LSMO initiates the elimination of La from the LSMO interfacial region, resulting in a more Sr rich alloy at the interface. This is well supported by the observed decrease in the XMCD intensity. Whereas La_0.7Sr_0.3MnO₃ is ferromagnetic, as the relative concentration of Sr is increased, the alloy becomes antiferromagnetic, no longer contributing to the XMCD signal. Therefore, although the XAS spectra (the average of the two helicity spectra, sensitive to the chemical state of the Mn) is continuously evolving with YBCO coverage, the XMCD spectra (the difference of the two helicity spectra, sensitive only to the ferromagnetic component of the Mn) simply diminishes in intensity without changing shape.

Figure 1: The Mn L-edge XAS (TOP) and XMCD (BOTTOM) spectra for various coverages of YBCO on LSMO. Included are spectra for 80 Å LNO/LSMO.

Figure 2: The Mn L-edge XAS and XMCD spectra for various coverages of STO (TOP) and LAO (BOTTOM) on LSMO.

The cation outdiffusion that we have identified here is not particular to the deposition of YBCO. In Figure 1, in addition to the YBCO coverage data, we have included the spectra for an 80 Å coverage of LNO (where similar interfacial degradation is observed). The LNO spectra match those for the evolving YBCO coverage quite well, suggesting that the interfacial disruption is not driven by the interfacial chemistry, but is most likely strain induced from lattice mismatch of the overlayer to the LSMO substrate.

Not all overlayer depositions result in interfacial degradation. The deposition of insulating STO and LAO on LSMO results in similar high quality epitaxial structures, but with very improved transport properties. A combined XAS and XMCD study of the interfacial development of these materials (shown in figure 2) show that no variations in the Mn L_2,3 edge XAS spectra occur and, for the STO overlayers, increased XMCD intensities are observed. The improved XMCD spectra indicates that the STO acts as an excellent passivation layer for LSMO and that the bare LSMO surface is slightly detrimentally affected by the in-air transfer between vacuum systems.

The elemental variation in interface distribution can also be quantified by X-ray resonant magnetic scattering (XRMS). XRMS is the angle dependent scattering of circular polarized X-rays, whose energy is tuned to the absorption edge of a magnetic element. Ydzerda and coworkers have shown that this method combines the element selectivity of X-ray resonant scattering with the magnetic contrast of XMCD and can be used to separately parameterize the magnetic and chemical roughness of buried interfaces. Such work on the oxide-oxide interfaces of our interest is currently in progress.

The work at the University of Maryland was supported by NSF MRSEC (Grant #DMR-96-32521) and ONR (Grant #N000149810092). The NRL work was supported by ONR. The Brookhaven National Laboratory is supported by DOE

Reference : The magnetism of a buried La0.7Sr0.3MnO3 interface., Stadler S, Idzerda YU, Chen Z, Ogale SB, Venkatesan T, Appl. Phys. Lett. 75, 3384 (1999).

---

Center for Superconductivity Research, University of Maryland, College Park, MD 20742-4111
Phone: 301.405.6129 Fax: 301.405.3779
Copyright © 2001 University of Maryland
Contact us with comments, questions and feedback